Oct 092015
 

On Wednesday Deutsche Bundesbank published “a list detailing its holdings of gold bars in custodian storage in Frankfurt am Main, London, Paris and New York”. While I welcome this move towards transparency and accountability on what is an important (and often controversial) public asset, it was disappointing that the Bundesbank did not think it necessary to produce a bar list that conformed to the usual standards that apply in the bullion market, which is to supply refinery, bar number, gross and fine weight, purity, and year of manufacture (where available).

The Bundesbank confirmed in an email to Ronan Manly that it has those details but said they were “not relevant for storage or accounting purposes”. They note that “whereas bar and melt numbers are issued by the gold refiner, inventory numbers are assigned by the custodian in whose vaults the gold is stored. Inventory numbers are either stamped or affixed to the bar using a label.” and identify different standards of disclosure between the four custodians of the bars:

  • Bank of England and the Banque de France: use internally assigned inventory numbers but are not willing to allow their customers to publish these
  • Bundesbank: uses internally assigned inventory numbers and has no problem publishing them
  • Federal Reserve Bank of New York: supplies the actual bar/melt number has not problem with customers publishing them

The Bank of England’s use of internally assigned inventory numbers (and only supplying that number, rather than the actual number stamped on the bar, to their custodian customers) was confirmed by Freedom of Information requests from blogger Bullion Baron in respect of Australia’s gold reserves.

I can understand that given the potential for duplicated bar numbers between refineries (and some refineries had a practise of restarting numbering each year) in any bar inventory control system it would be easier to design the database with its own unique identification key rather than rely on the combination of refiner/number/weight/purity/year as the unique key. However, this does not mean a custodian needs to stick a label on each bar as bars can be located and identified by their vault location (ie shelf or pallet number) and unique bars markings if needed. The Perth Mint manages to keep track of its physical stocks and client Allocated without needing sticky labels, so it is possible.

While internally assigned inventory numbers might suffice for “storage or accounting purposes” I would have thought it obvious that they do not suffice for independent audit purposes. I doubt any auditor would accept a sticky label on a bar for the simple fact that a custodian could stick different numbered labels on the same bar each time different customers come around to audit. This is why all of the exchange traded funds that publish bar lists provide full bar list details and do not list any internally assigned inventory numbers. This is clearly the industry standard and odd why the Bundesbank could not have just supplied the additional information that they have in their spreadsheets.


On to the geeky stuff of looking at the bar list itself. It would have been helpful if the Bundesbank supplied the bar list in spreadsheet form, rather than pdf, but thankfully the bar list guru, Warren James (who has a massive database of the major ETF bar lists) was able to convert the Bundesbank pdf into a 13mb spreadsheet, which you can download if you so desire.

A review of the pdf list shows that it has not been sorted by internally assigned inventory number and that some other sorting has been applied, most likely pallet or some other location identifier, or possibly by refiner, given the grouping we see in some of the Federal Reserve bar numbers.

One of the first things I noticed when doing a simple sort by weight check is bar number 11202, which supposedly weighs 23.810 ounces (see extract from bar list below from page 2258):

buba1

There are eight other circa 200oz bars (which are odd in themselves) so I would guess this is a keying error and should be 238.100 ounces. A 23 ounce bar seems highly unlikely but maybe in the olden days if the US Mint was doing a melt and had an amount of gold left over that wasn’t near 400oz they just poured whatever remained. Or maybe it is some sort of Easter Egg that the FRB NY put in to see if the Bundesbank was actually reviewing their bar list. Might be worth a call to the FRB NY to get them to check this bar and confirm its actual weight.

The US has previously released it bar lists (see here) and it gives us something to compare to, particularly since some of Germany’s gold is held with the FRB NY. In contrast to the US’ gold, almost all of Germany’s gold is within the LBMA London Good Delivery specification weight of 350-430oz. Note that German holds a total of 270,316 bars versus the US’ 699,515 bars.

Weight Germany US
<350oz 0% 37%
LBMA Standard 350oz-430oz 99% 54%
430oz-452oz 1% 9%
Total 100.0% 100.0%

In terms of purity, all of Germany’s gold meets the minimum London Good Delivery specification of 99.50% or higher.

Purity Germany US
<99.50% 0% 83%
99.50% to 99.89% 85% 3%
99.90% to 99.98% 11% 10%
99.99%+ 3% 4%
Total 100% 100%

The large number of US sub-purity bars reflects coin melt from the 1930s confiscation. What is interesting is the 3 to 4% which are 99.99% or higher, including the existence of some rare 99.999% (five nines), as you can see from this extract from the bar list (these bars being held at the FRB NY):
buba2

The out of the 9,273 99.99%+ purity bars the Bundesbank holds 3,002 (totalling 1,197,116.298 ounces) are in the Bank of England and 3351 (1,326,074.084 ounces) are held by the Bundesbank itself. The reason I point this out is that when Asian demand is high, 99.99%+ purity 400oz bars can attract a premium up to $0.50 an ounce as refiners only need to melt and recast them into kilobars, which gives a lot quicker turnaround than refining dore up to 99.99%.

With over 2.5 million ounces of 99.99% bars easily shipped to Swiss refiners, the Bundesbank could earn over $1 million dollars by swapping them for 99.50% purity bars the next time Chinese demand surges (less freight costs). There might be some work getting all the 99.99% bars together but at least the bars in the picture below (from Koos Jansen’s post on the Bundesbank bar list) are nicely grouped together:

You can match the sticky label internally generated numbers above to the bar list below to see they are all 99.99%buba3

Just out of interest, the US holds 24,946 99.99%+ bars, which at $0.50 premium could be worth $5 million on a swap. Seems such a waste having these high purity bars just sitting in central banks vaults when the Chinese would love to have them.